DEV Community

Arin Kishore
Arin Kishore

Posted on

From 0 to 100% API Test Coverage with Keploy AI – My Journey

Over the past few days, I had the amazing opportunity to work on API testing using Keploy AI as part of the Keploy API Fellowship. In this post, I’ll walk you through everything I did — from building a Node.js API to running AI-powered tests and integrating it into a CI/CD pipeline.

🔧 My API Project
I built a Student Manager API using:

  1. Node.js & Express – for the backend
  2. MongoDB Atlas – as the database
  3. Swagger – to document the API
  4. Endpoints for GET, POST, PUT, DELETE at /api/students

GitHub Repository:
https://github.com/kishorecodesinpython/student-api-server

🧪 Task 1 – API Testing with Keploy AI
Step 1: Created an OpenAPI Schema

I defined all endpoints and schemas using Swagger UI, hosted at /api-docs.

Step 2: Ran Keploy in Docker

Since I’m using Windows, I had to use Docker with WSL2. I ran this command: docker compose up --build
This built and launched my API and Keploy CLI together inside containers.

Step 3: Recorded API Calls

I sent multiple requests using curl and Postman to record traffic, while Keploy captured them in real-time. Then I ran: keploy test ...
This generated multiple test cases from actual traffic. I got a Test Drive report with:

  • 27 Test Suites
  • 20 Accepted
  • 7 Rejected

Step 4: Debugging Docker & Environment Issues

This was not all smooth! I faced a few problems:

  • Docker WSL2 was broken (resolved via reset and reinstall)
  • MongoDB URI wasn’t passed properly (fixed using dotenv)
  • Curl commands needed to be corrected for schema match

Step 5: CI/CD Integration

I integrated Keploy testing into a GitHub Actions pipeline, which automatically:

  • Built my app using Docker
  • Ran all tests
  • Validated test outputs

🌐 Task 2 – Chrome Extension API Testing
I explored the Keploy Chrome Extension to test real-world APIs.

Site 1: DummyJSON
I captured a GET request to /products using the Chrome console and the Keploy extension.

Site 2: JSONPlaceholder
Tested endpoints like GET /posts, POST /posts, and validated response handling.

The Chrome Extension made it incredibly easy to record calls and generate test cases on the fly.

💡 What I Learned

  • Keploy’s AI-generated tests helped me go from zero to complete test coverage in minutes.
  • Docker with WSL2 on Windows takes patience and careful setup.
  • The Chrome extension is perfect for testing third-party/public APIs.
  • CI/CD test integration adds confidence to production readiness.

📸 Final Screenshots I Shared:

  • Swagger API Docs UI
  • Keploy “Test Drive” bunny report
  • Docker logs running Keploy
  • MongoDB connected confirmation in terminal

🏁 Conclusion
Thanks to Keploy, I transitioned from writing tests manually to using AI for full automation. This fellowship was one of the most hands-on testing experiences I’ve had — and I’ll definitely be applying these workflows to future projects.

GitHub Repo:
https://github.com/kishorecodesinpython/student-api-server

Let me know what you think or if you want to connect!

Keploy #APITesting #CI_CD #Nodejs #MongoDB #Docker #OpenSource #AIinTesting #KeployFellowship

Image description
Image description
Image description
Image description

Top comments (0)